JP6678357B2 - Motion vector selection and prediction method in video coding system - Google Patents

Description

  The present invention relates to encoding and decoding video signals, and more particularly, to selecting predicted motion vectors for frames of a video sequence.

  With the advent of digital multimedia, such as digital images, speech / speech, graphics, and video, a variety of applications, including new and published applications, have been significantly improved to provide reliable storage, communication, transmission, content search, Access is relatively easy. Overall, digital multimedia applications cover a wide range of fields, including entertainment, information, healthcare, and security, and are useful to society in many ways. Multimedia captured by sensors such as cameras and microphones is often analog and digitized by a digitizing process in PCM (Pulse Coded Modulation) format. However, after digitization, the amount of data needed to recreate the analog representation required for speakers and / or TV displays becomes important. That is, efficient communication, storage or transmission of large amounts of digital multimedia content requires compression from the raw PCM format to a compressed representation. Accordingly, many techniques have been invented for multimedia compression. Over the years, video compression technology has grown very high in achieving high compression factors between 10 and 100, while maintaining high psychological visual quality, similar to uncompressed digital video. Was.

  (Video coding standards led by a number of standards bodies such as MPEG-1, MPEG-2, H.263, MPEG-4 part2, MPEG-4 AVC / H.264, MPEG-4 SVC and MVC, Windows Media To date, enormous advances have been made in video compression technology and science, as indicated by industry-leading proprietary standards such as Video, RealVideo, and On2VP, but with higher quality and higher definition. Increasing consumer demand to access 3D (stereo) anytime, anywhere, and now, through various means such as DVD / BD, radio broadcasting, cable / satellite broadcasting, wired and mobile networks, A higher level of video delivery to a variety of client devices, such as PCs / laptops, televisions, set-top boxes, game consoles, portable media players / devices, smartphones, wearable computing devices, etc. There is a need to deliver by condensation. For standards led by standards bodies, ISO MPEG's recently launched efforts in high-efficiency video coding have contributed new technologies and technologies from years of research into H.265 video compression by the ITU-T standards committee, It became clear that the combination of is expected.

  All of the above standards employ a general inter-frame predictive coding structure that includes reducing temporal redundancy by compensating for motion between video frames. The basic idea is to remove the time dependency between adjacent pictures using a block matching method. At the beginning of the encoding process, each frame of the unencoded video sequence is divided into three categories: I-type frames, P-type frames, and B-type frames. Type I frames are inner coded. That is, it contains only information of frames that are themselves used to encode the image and do not use inter-frame motion compensation techniques (intra-frame motion compensation techniques may be applied).

  The other two types of frames, P-type and B-type, are encoded using inter-frame motion compensation techniques. The difference between a P picture and a B picture is the temporal direction of a reference picture used for motion compensation. P-type pictures use information from previous pictures in display order, while B-type pictures use information from previous and future pictures in display order.

For P-type and B-type frames, each frame is divided into blocks of pixels represented by the coefficients of the luminance and chrominance signal components of each pixel, and each block has one or more motion vectors (for B-type pictures, And two motion vectors may be encoded for each block, since information from previous encoded frames can be utilized). A motion vector (MV) represents a spatial displacement from a position in the current block to a position in a similar block in another, previously encoded frame (which may be a past or future frame in display order). , Respectively, including a reference block and a reference frame. To generate a residual (also called a "residual signal"), the difference between the reference block and the current block is calculated. Therefore, for each block of the inter-coded frame , it is necessary to code only the residuals and motion vectors, not the entire contents of the block. By removing such temporal redundancy between frames of the video sequence, the video sequence can be compressed.

  After intra- or inter-frame prediction techniques have been applied to further compress the video data, the coefficients of the residual signal are often calculated using the discrete cosine transform (DCT) or discrete sine transform (DST). ,) From the spatial domain to the frequency domain. In the case of naturally occurring images, such as the types of images that typically make up a human perceptible video sequence, low frequency energy is always stronger than high frequency energy. Therefore, the residual signal in the frequency domain can perform better energy compression than in the spatial domain. After the forward transform, the coefficients and motion vectors can be quantized and entropy coded.

  On the decoding side, inverse quantization and inverse transform are applied to recover the spatial residual signal. These are common transformation / quantization steps in all video compression standards. An inverse prediction step is performed to generate a reproduced version of the original unencoded video sequence.

  In past standards, the blocks used for encoding were typically 16x16 pixels (called macroblocks in many video encoding standards). However, since the development of these standards, frame sizes have become larger and many devices have higher display capabilities than "high definition" (or "HD") frame sizes, such as 2048 x 1530 pixels. . Therefore, in order to efficiently encode these frame size motion vectors, it is desirable to have larger blocks, such as 64x64 pixels. On the other hand, it is desirable to be able to perform motion prediction and conversion on a relatively small scale, for example, 4 × 4 pixels, for a corresponding improvement in resolution.

  As the resolution of motion estimation increases, the amount of bandwidth required to encode and transmit motion vectors, frame by frame, and optionally over a video sequence, increases.

FIG. 1 illustrates an example video encoding / decoding system in at least one embodiment. FIG. 4 is a diagram exemplarily illustrating components of an exemplary encoding device in at least one embodiment. FIG. 7 is a diagram exemplarily illustrating components of an exemplary decoding device in at least one embodiment. FIG. 4 is a block diagram of an example video encoder in at least one embodiment. FIG. 9 is an example block diagram of an example video decoder in at least one embodiment. FIG. 7 is a diagram exemplarily illustrating an exemplary motion vector selection routine in at least one embodiment. FIG. 7 is a diagram exemplarily illustrating an exemplary motion vector candidate generation subroutine in at least one embodiment. FIG. 4 is a diagram exemplarily illustrating an exemplary motion vector restoration routine in at least one embodiment.

  The following detailed description primarily illustrates the operation steps and symbolic representations of conventional computer components, including the processing device, the memory storage of the processing device, the connected display device, and the input device. Further, these steps and representations can utilize conventional computer components in a heterogeneous distributed computer environment, including remote file servers, computer servers, and memory storage devices. Each of these conventional distributed computer components is accessible by a processor over a communications network.

  The phrases “in one embodiment,” “in at least one embodiment,” “in various embodiments,” “in some embodiments,” and the like are used repeatedly herein. Such phrases are not necessarily referring to the same embodiment. The terms "comprising," "having," and "including" are synonymous, unless the context indicates otherwise. The various embodiments describe a general "hybrid" video coding approach used in intra / inter-picture prediction and transform coding, as generally described above.

  Attention is now directed to the detailed description of the embodiments illustrated in the drawings. Embodiments are described with reference to the drawings and the associated description, but those skilled in the art may determine whether they are explicitly shown and / or described without departing from the scope of the present disclosure. However, it can be appreciated that instead of the specific embodiment shown or described, alternative and / or equivalent embodiments including all alternatives, modifications, and equivalents are possible. . Without limiting the scope to the embodiments disclosed herein, various different embodiments, additional devices, or combinations of the illustrated devices can be added or combined. In alternative embodiments, additional devices or combinations of the illustrated devices can be added or combined without limiting the scope to the embodiments disclosed herein.

Exemplary Video Encoding / Decoding System FIG. 1 illustrates an exemplary video encoding / decoding system 100 in at least one embodiment. An encoding device 200 (shown in FIG. 2 and described below) and a decoding device 300 (shown in FIG. 3 and described below) communicate data using the network 104. Decryption device 200 may be either a direct data connection via a storage area network (“SAN”), a high speed serial bus, and / or other suitable communication technology, or network 104 (shown in dashed lines in FIG. 1). Through one, it is in data communication with an unencoded video source 108. Similarly, the decryption device 300 may be a direct data connection via a storage area network ("SAN"), a high speed serial bus, and / or other suitable communication technology, or the network 104 (shown in dashed lines in FIG. 1). In data communication with any encoded video source 112. In some embodiments, encoding device 200, decoding device 300, encoded video source 112, and / or unencoded video source 108 may include one or more replicated and / or distributed It has physical or logical devices. In many embodiments, there may be more encoding devices 200, decoding devices 300, unencoded video sources 108, and / or encoded video sources 112 than shown.

  In various embodiments, encoding device 200 is generally a networked computer device that accepts requests, for example, from decoding device 300 over network 104 and provides a response accordingly. In various embodiments, the decoding device 300 includes a mobile phone, a watch, glasses, or other wearable computing device, a dedicated media player, a computer tablet, a car head, an audio-video on demand (AVOD) system, a dedicated media console, It may be a networked computer device having external components such as a game device, a "set-top box", a digital video recorder, a television, or a general-purpose computer. In various embodiments, network 104 may include the Internet, one or more local area networks (“LANs”), one or more wide area networks (“WANs”), a cellular data network, and / or Including data networks. Network 104 can be a wired and / or wireless network in various respects.

Exemplary Encoding Device FIG. 2 shows components of an exemplary encoding device 200. In some embodiments, the encoding device includes more components than those shown in FIG. However, not all of these general conventional components need be shown to disclose the illustrative embodiments. As shown in FIG. 2, the exemplary encoding device 200 includes a network interface 204 that connects to a network, such as the network 104. The example encoding device 200 also includes a processing device 208, a memory 212, an optional user input device 214 (eg, an alphanumeric keyboard, keypad, mouse or other pointing device, touch screen, and / or microphone) and any It includes a display 216, all of which are interconnected with the network interface 204 via a bus 220. Memory 212 typically includes RAM, ROM, and permanent mass storage such as disk drives, flash memory, and the like.

  The memory 212 of the exemplary encoder 200 includes an operating system 224 and an inter-frame video encoder implemented by software to implement the motion vector selection routine 600 (described below with reference to FIG. 6). A number of software service program codes, such as 400 (described below with reference to FIG. 4). The memory 212 also comprises a video data file (not shown) representing an unencoded copy of the audio / visual media work, such as a movie and / or television series. These and other software components are encoded using a drive (not shown) associated with a non-transitory computer readable medium 232, such as a floppy disk, tape, DVD / CD-ROM drive, memory card, or the like. May be read into the memory 212 of the conversion apparatus 200. Although the exemplary encoding device 200 has been described, the encoding device can communicate with the network 120 and implement video coding, such as exemplary software or a motion vector selection routine 600 implemented by the video encoding unit 400. The computer device may be any of a number of networked computer devices capable of executing instructions that implement software.

  In operation, the operating system 224 manages the hardware and other software resources of the encoding device 200 and provides general services of software applications such as the inter-frame video encoder 400 implemented by software. Memory 212 for various software applications such as network communication via the network interface 204, data reception via the input device 214, data output via the display 216, software-implemented inter-frame video encoder 400, etc. Due to hardware functions such as assignment of the operating system 224, the operating system 224 acts as an intermediary between software and hardware implemented by the encoding device.

  In some embodiments, the encoding device 200 further includes a dedicated unencoded video interface 236 that communicates with the unencoded video source 108, for example, a high-speed serial bus. In some embodiments, encoding device 200 communicates with unencoded video source 108 via network interface 204. In other embodiments, unencoded video source 108 resides in memory 212 or computer readable media 232.

  Although the encoding device 200 has been described with reference to an exemplary encoding device 200 that generally conforms to a conventional general-purpose computer device, the encoding device 200 may include, for example, a video recording device, a video processing and / or accelerator, a personal computer, and a game main unit. , A set-top box, a personal digital assistant or a wearable computer device, a smart phone, or any other suitable device.

  Encoding device 200 may be used, for example, for an on-demand media service (not shown). In at least one exemplary embodiment, an on-demand media service encodes for facilitating an on-demand media store that provides digital copies of media works, such as video content, to users on a per-works and / or subscription basis. The device 200 is managed. The on-demand media service obtains digital copies of such media works from unencoded video sources 108.

Exemplary Decoding Device FIG. 3 shows components of an exemplary decoding device 300. In some embodiments, the decoding device includes more components than those shown in FIG. However, not all of these general conventional components need be shown to disclose the illustrative embodiments. As shown in FIG. 3, the exemplary decoding device 300 includes a network interface 304 that connects to a network, such as the network 104. The example decoding device 300 also includes a processing device 308, a memory 312, an optional user input device 314 (eg, an alphanumeric keyboard, keypad, mouse or other pointing device, touch screen, and / or microphone) and any It includes a display 316 and optional speakers 318, all interconnected with a network interface 304 via a bus 320. Memory 312 typically includes RAM, ROM, and permanent mass storage such as disk drives, flash memory, and the like.

  The memory 312 of the exemplary decoding device 300 includes an operating system 224 and a video decoding unit 500 (implemented by software to implement a motion vector restoration routine 800 (described below with reference to FIG. 8)). (Described below with reference to FIG. 5). The memory 312 also includes a video data file (not shown) that represents an encoded copy of the audio / visual media work, such as a movie and / or television series. These and other software components are decrypted using a drive (not shown) associated with a non-transitory computer readable medium 332 such as a floppy disk, tape, DVD / CD-ROM drive, memory card, or the like. May be read into the memory 312 of the optimization device 300. Although the exemplary decoding device 300 has been described, the decoding device may communicate with the network 120 and may implement a number of example software or message extraction routines 700 implemented by the video decoding unit 500, etc. It may be some computer device of a networked computer device group.

  In operation, operating system 324 manages the hardware and other software resources of decoding device 300 and provides general services for software applications, such as video decoding unit 500, implemented by software. For hardware functions such as network communication via the network interface 304, data reception via the input device 314, data output via the display 316 or any speaker 318, allocation of memory 312, the operating system 324 Functions as an intermediary between software and hardware implemented by the encryption device.

  In some embodiments, the decoding device 300 further includes an optional encoded video interface 336 that communicates with the encoded video source 116, for example, a high-speed serial bus. In some embodiments, the decoding device 300 communicates with the encoded video source 116 via the network interface 304. In other embodiments, encoded video source 116 resides in memory 312 or computer readable media 332.

  Although the decoding device 300 has been described as an exemplary decoding device 300 that generally conforms to a conventional general-purpose computer device, the decoding device 300 may be, for example, a video recording device, a video processing and / or accelerator, a personal computer, a game console, or the like. , A set-top box, a personal digital assistant or a wearable computer device, a smart phone, or any other suitable device.

  The decoding device 300 may be used, for example, for an on-demand media service. In at least one exemplary embodiment, the on-demand media service provides digital copies of media works, such as video content, on a work-by-work and / or subscription basis to users operating the decryption device 300. The decoding device obtains a digital copy of such media work via the network 104 from an uncoded video source 108, such as the coding device 200, for example.

FIG. 4 illustrates an inter-frame video encoder 400 (hereinafter “code”) that employs an enhanced motion vector selection and prediction technique according to at least one embodiment and is implemented by software. A general functional block diagram of the “modification unit 400” is shown. One or more uncoded video frames (vidfrms) in the video sequence are provided to sequencer 404 in display order.

  The sequencer 404 assigns a predictive coded picture type (e.g., I, P or B) to each uncoded video frame, and reorders the frames or frames in a frame array into a coding order (e.g., I- Rearrange according to a type frame, followed by a P type frame, followed by a B type frame). The ordered uncoded video frames (seqfrms) are input to the block indexing unit 408 in the coding order.

  For each ordered uncoded video frame (seqfrms), block indexing unit 408 determines the largest coded block (“LCB”) size of the current frame (eg, 64 × 64 pixels), and The unencoded frames are divided into columns of encoded blocks (cblks). The coding blocks in the frame have different sizes, for example, the LCB size in the current frame is 4 × 4 pixels.

  Each coded block is input to the difference unit 412 one at a time, and the difference from the corresponding prediction signal block (pred) generated from the previously coded block is calculated. The coded blocks (cblks) are provided to a motion estimator 416 (described below). After the difference is calculated by the difference unit 412, the obtained residual signal (res) is forward-transformed into a frequency domain expression by the conversion unit 420, and a block of the conversion coefficient (tcof) is obtained. Next, the block of the transform coefficient (tcof) is sent to the quantization unit 424, and a block of the quantization coefficient (qcf) is generated and sent to both the entropy coding unit 428 and the local decoding loop 430.

  At the start of the local decoding loop 430, the inverse quantization unit 432 inversely quantizes the block of transform coefficients (tcof ') and generates them by the inverse transformation unit to generate an inverse quantization residual block (res'). Pass to 436. In the adding unit 440, the prediction block (pred) from the motion compensation prediction unit 442 is added to the inverse quantization residual block (res ′) in order to generate a locally decoded block (rec). The locally decoded block (rec) is sent to the frame assembling unit and the deblocking filter processing unit 444, where the blocking is reduced and the restored frame (recd) is assembled. Used as a reference frame.

  The entropy encoding unit 428 encodes the quantized transform coefficient (qcf), the differential motion vector (dmv), and other data, and generates an encoded video bit stream 448. For each frame of the uncoded video sequence, the coded video bitstream 448 may include coded picture data (eg, coded and quantized transform coefficients (qcf) and differential motion vectors (dmv). )) And an encoded frame header (eg, syntax information such as the LCB size of the current frame).

According to the function of the motion estimator 416, as described in detail below with reference to FIGS. 6 and 7, according to at least one embodiment, each coded block in a P-type or B-type frame is: , residual in the current encoding block is generated based on the same frame previously coded blocks in the (all coded blocks as if it is I-type) ( "reference block"), or, inter-coded , i.e., the residual of the current encoding block is generated based on one or more reference blocks from one or more previous encoded encoded frame (the "reference frame"). In the illustrated embodiment, the inter-encoded block is processed in any suitable manner.

According to at least one embodiment, for an inter-coded coded block, the motion estimator 416 determines at least three motion vectors for the coded block (or group of B-type frames). One of the three coding modes is selected: a skip coding mode, a direct coding mode, and an inter coding mode .

Inter coding mode
For coding blocks that are coded in the inter coding mode , the motion estimator 416 converts each coding block into, for example, 4 × 4 pixels, 8 × 8 pixels, 16 × 16 pixels, 32 × 32 pixels, or 64 × Divide into 64 pixel prediction blocks. For example, a 64 × 64 coded block is divided into 16 16 × 16 prediction blocks, 4 32 × 32 prediction blocks, or 2 32 × 32 prediction blocks and 8 16 × 16 prediction blocks. can do. Next, the motion estimator 416 calculates a motion vector (MV calc) for each prediction block by determining an appropriate reference block and determining a relative spatial displacement from the prediction block to the reference block.

According to one aspect of at least one embodiment, to increase coding efficiency, the calculated motion vector (MV calc) is calculated from the calculated motion vector (MV calc) by using a motion vector predictor (MV pred). ) To generate a motion vector difference (ΔMV). For example, if the calculated motion vector (MV calc) is (5, -1) (ie, the reference block from the previously encoded frame is 5 columns to the right and 1 column to the current prediction block in the current frame). And the motion vector predictor is (5,0) (ie, the reference block from the previously encoded frame is 5 columns relative to the current prediction block in the current frame). (Located on the right and in the same row), the motion vector difference (ΔMV) is as follows:
MV calc−MV pred = (5, -1) − (5,0) = (0, -1) = ΔMV

  The closer the motion vector predictor (MV pred) is to the calculated motion vector (MV calc), the smaller the value of the motion vector difference (ΔMV). Thus, by accurately predicting motion vectors and encoding the motion vector differences instead of the calculated motion vectors, much less information is needed throughout the video sequence. Note that this prediction is irrelevant to the content of the current prediction block and is reproducible on the multifunction peripheral side.

  According to one aspect of at least one embodiment, motion estimator 416 may use various techniques to obtain a motion vector predictor (MV pred). For example, a motion vector predictor can be obtained by calculating an intermediate value of a plurality of previously encoded motion vectors in a prediction block of a current frame. Also, for example, a motion vector predictor is a motion vector in a reference block (RBa) located in the same column and one row above the current block, or one example right and one row below the current prediction block. , And the motion vector of the reference block (RB b) located on the left side of the current block and the reference block (RB c) in the same row as the current block, which have been previously encoded in the spatial neighborhood of the current prediction block It may be an intermediate value of a plurality of reference blocks.

As described above, according to an aspect of at least one embodiment, the motion estimator 416 uses additional or alternative techniques to provide a motion vector predictor for a predicted block in an inter-coding mode . be able to. For example, in other techniques for providing a motion vector predictor, a motion vector in a reference block (RBa) located on the same column and one row for the current block, or one example for the current prediction block And the motion vector of the reference block (RB b) located one row below, and the previous one in the spatial neighborhood of the current prediction block, like the reference block (RB c) one column to the left of the current block and the same row The average value of a plurality of encoded reference blocks is determined.

  According to an aspect of at least one embodiment, to increase coding efficiency, the coding unit 400 may include a motion vector selected in a picture header of a current frame (or a prediction block header of a current prediction block). Placing a prediction method (SMV-PM) flag may indicate which of the available techniques was used to encode the current prediction block. For example, in at least one embodiment, the SMV-PM flag is one bit with two possible values, where one possible value is obtained when the motion vector predictor is obtained using the intermediate value technique described above. And the second possible value indicates that the motion vector predictor was obtained using an alternative technique.

In a coded block coded according to the inter coding mode , both the motion vector and the residual are coded and included in the bit stream.

Skip coding mode and direct coding mode In the case of a coded block coded in the skip coding mode or the direct coding mode, the motion estimator 416 uses the entire coded block as a corresponding prediction block (PB). .

  According to an aspect of at least one embodiment, in skip coding and direct coding modes, rather than determining a motion vector (MV calc) calculated for a predicted block (PB), a motion estimator 416. Uses a predetermined method described below with reference to FIG. 7 to generate an ordered list of motion vector candidates. For example, for the current prediction block (PB cur), an ordered list of motion vector candidates may be used previously to encode another block of the current frame, called "reference blocks" (RBs). It consists of the motion vector obtained.

  According to an aspect of at least one embodiment, the motion estimator 416 selects the best candidate motion vector (MVC) from an ordered list that encodes the current prediction block (PB cur). If the step of generating an ordered list of motion vector candidates is reproducible on the decoding device side, only the index of the selected motion vector (MV sel) in the ordered list of motion vector candidates is not the motion vector itself. It may be included in the encoded bit stream. By encoding the index values rather than the actual motion vectors, much less information is needed throughout the video sequence.

  According to an aspect of at least one embodiment, the motion vectors selected to generate the motion vector candidate list are preferably three reference blocks (RBa, RBb, RBc) having known motion vectors. ) And shared by the current prediction block (PB cur) and / or other reference blocks (RB). For example, the first reference block (RB a) is located directly above the current prediction block (PB cur), and the second reference block (RB b) is located to the right of the first reference block (RB a). However, the third reference block (RB c) is located to the left of the current prediction block (RB c). However, the specific position of the reference block with respect to the current prediction block is not particularly important as long as they are predefined and it is possible to know where the downstream decoding device is located.

  According to an aspect of at least one embodiment, if all three reference blocks have known motion vectors, the first motion vector candidate (MVC 1) in the motion vector candidate list of the current prediction block (PB cur). ) Is a motion vector (MVa) from the first reference block (RB a) (a group of motion vectors in the B-type frame), and the second motion vector candidate (MVC 2) is a second reference block (MVC 2). RB b) (MV b) (or a group of motion vectors) from the third reference block (RB c), and the motion vector (MV c) from the third reference block (RB c). Or a group of motion vectors). Therefore, the motion vector candidate list is (MVa, MVb, MVc).

  However, for example, either because the prediction information is not available for a particular reference block, or because the current prediction block (PB cur) is located in the top row, left end row, or right end row of the current frame, etc. If one of the reference blocks (RBs) does not have an available motion vector, the motion vector candidate is omitted, the next motion vector candidate is located at that position, and the zero-value motion vector (0,0) is the remaining candidate level. Be substituted as For example, when a motion vector is not available for RBb, the motion vector candidate list is (MVa, MVc, (0,0)).

In at least one embodiment, the full set of combinations of motion vector candidate lists indicating valid combinations of motion vector candidates is as shown in Table 1.

  Next, the motion estimator 416 evaluates the motion vector candidates and selects the best motion vector candidate to be used as the selected motion vector for the current prediction block. If the downstream decoding device knows how the ordered list of motion vector candidates for the prediction block has been generated, this calculation is reproduced on the decoding device side that does not know the content of the current prediction block. Note that it is possible. Therefore, only the index of the motion vector selected from the motion vector candidate list, for example, by setting a motion vector selection flag in the prediction block header of the current prediction block, instead of the motion vector itself, is encoded. By encoding the index values that need to be included in the bitstream and not the actual motion vectors, much less information is needed throughout the video sequence.

  In the direct encoding mode, the motion vector selection flag and the residual between the current prediction block and the block of the reference frame indicated by the motion vector are encoded. In the skip encoding mode, the motion vector selection flag is encoded, but encoding of the residual signal is omitted. In essence, this indicates instructing the downstream decoding device to use the block of the reference frame indicated by the motion vector instead of the current prediction block of the current frame.

FIG. 5 illustrates an inter-frame video adapted for use with a decoding device, such as decoding device 300, employing an improved motion compensated prediction technique according to at least one embodiment. FIG. 4 shows a general functional block diagram of corresponding software implemented by a decoding unit 500 (hereinafter, referred to as “decoding unit 500”). Decoding section 500 operates in the same manner as local decoding loop 455 in encoding section 400.

  Specifically, the encoded video bitstream 504 to be decoded includes quantization coefficients (qcf), differential motion vectors (dmv), accompanying message data packets (msg-data), and other blocks of data. Provided to an entropy decoding unit 508 that can decode. Next, the block of the quantization coefficient (qcf) is inversely quantized by the inverse quantization unit 512, and an inverse quantization coefficient (tcof ′) is obtained. Next, the inverse quantization coefficient (tcof ′) is inversely transformed from the frequency domain by the inverse transform unit 516 to obtain a decoded residual block (res ′). The adding unit 520 adds the motion-compensated prediction block (pred) obtained using the corresponding motion vector (mv). The resulting decoded video (dv) is subjected to deblocking filtering in a frame assembly unit and a deblocking filtering unit 524. The blocks (recd) at the output of the frame assembler and deblocking filter processing unit 528 form a reconstructed frame of the video sequence, are output from the decoding unit 500, and are motion compensated for decoding subsequent coded blocks. Used as a reference frame for predictor 532.

Motion Vector Selection Routine FIG. 6 shows a motion vector selection routine 600 suitable for use with at least one embodiment, such as the encoder 400. As will be appreciated by those skilled in the art, not all events in the encoding process are shown in FIG. Rather, for clarity, only those steps that are reasonably relevant to describing the motion vector selection routine are shown.

  In the execution block 603, for example, the motion estimation unit 416 acquires an encoded block.

At decision block 624, the motion vector selection routine 600 selects an encoding mode for the encoded block. For example, as described above, the inter encoding mode , the direct encoding mode, or the skip encoding mode is selected. If either the skip encoding mode or the direct encoding mode has been selected for the current encoded block, the motion vector selection routine 600 proceeds to an implementation block 663 described below.

If, at decision block 624, the inter-coding mode has been selected for the current coded block, then at implementation block 627, the motion vector selection routine 600 converts the current coded block to one or more prediction blocks. Along with splitting, a start loop block 630 is started, where each prediction block in the current coding block is processed in turn.

  In implementation block 633, the motion vector selection routine 600, for a B-type picture, indicates whether the reference frame is a previous picture, a future picture, or both, of the current prediction block. Select a prediction index.

  Next, at implementation block 636, the motion vector selection routine 600 selects a motion vector prediction method, such as the median or average technique described above, or any available alternative motion vector prediction method.

  At an implementation block 642, the motion vector selection routine 600 obtains a motion vector predictor (MV pred) for the current prediction block using the selected motion vector prediction method.

  In an implementation block 645, the motion vector selection routine 600 obtains a calculated motion vector (MV calc) for the current prediction block.

  In an implementation block 648, the motion vector selection routine 600 determines the motion vector difference (ΔMV) of the current prediction block (a single motion vector difference for a P-type picture, and two motion vectors for a B-type picture). Difference exists).

  In the implementation block 651, the motion vector selection routine 600 obtains the residual between the block indicated by the calculated motion vector (MV calc) and the current prediction block (PB cur).

  At implementation block 654, the motion vector selection routine 600 encodes the motion vector differences and residuals of the current prediction block.

  In an implementation block 657, the motion vector selection routine 600 sets the SMV-PM flag indicating which motion vector prediction technique was used for the current prediction block, in the picture header of the current frame (or in the current prediction block). (Predicted block header).

  At the end loop block 660, the motion vector selection routine 600 returns to the start loop block 630 to process the next prediction block (if any) of the current coded block.

  Returning to decision block 624, if either the skip coding mode or the direct coding mode has been selected for the current coded block, then in implementation block 663, the motion vector selection routine 600 returns to the current coded block. Set the current prediction block to be equal to

  Next, the motion vector selection routine 600 calls a motion vector candidate generation subroutine 700 (described below with reference to FIG. 7) that returns an ordered list of motion vector candidates to the motion vector selection routine 600.

  Next, in an implementation block 666, the motion vector selection routine 600 selects a motion vector from the motion vector candidate list for use in encoding the current prediction block.

  If, at decision block 667, the selected coding mode is direct coding, then at implementation block 669, the motion vector selection routine 600 determines between the current prediction block and the reference block indicated by the selected motion vector. Calculate the residual of.

  In an implementation block 672, the motion vector selection routine 600 encodes the residual, and in an implementation block 675, the motion vector selection routine 600 determines which motion vector candidate was selected for encoding the current prediction block. A vector selection flag is set in the prediction block header of the current prediction block.

  The motion vector selection routine 600 ends at end block 699.

Motion vector candidate generation subroutine 700
FIG. 7 illustrates a motion vector candidate generation subroutine 700 for generating an ordered list of motion vector candidates according to at least one embodiment. In the illustrated embodiment, three motion vector candidates are generated. However, one of ordinary skill in the art will recognize that more or lesser amounts of candidates may be generated using the same technique, and further alternative and / or equivalent without departing from the scope of the present disclosure. It can be appreciated that the embodiments of the present invention are substitutable for the particular embodiment shown and described.

  The motion vector candidate generation subroutine 700 obtains a request to generate a motion vector candidate list of a current prediction block in an execution block 704.

  If, at decision block 708, a motion vector is available from the first candidate reference block (RBa), at implementation block 712, the motion vector candidate generation subroutine 700 converts the first motion vector candidate (MVC 1) to an MV Set to a and proceed to decision block 716.

  If, at decision block 716, a motion vector is available from the second candidate reference block (RB b), then at execution block 724, the motion vector candidate generation subroutine 700 converts the second motion vector candidate (MVC 2) to an MV Set to b and proceed to decision block 728.

  If, at decision block 728, a motion vector is available from the third candidate block (RBc), then at implementation block 736, the motion vector candidate generation subroutine 700 converts the third motion vector candidate (MVC3) to MVc. Set to.

  Next, the motion vector candidate generation subroutine 700 returns a motion vector candidate list having respective values of MVC 1 = MVa, MVC 2 = MVb, and MVC 3 = MVc in a return block 799.

  Referring again to decision block 728, if a motion vector is not available from the third candidate block (RBc), then at execution block 740, the motion vector candidate generation subroutine 700 converts the third motion vector candidate (MVC3) to ( 0,0).

  Next, the motion vector candidate generation subroutine 700 returns a motion vector candidate list having respective values of MVC 1 = MVa, MVC 2 = MVb, and MVC 3 = (0,0) in a return block 799.

  Referring again to decision block 716, if a motion vector is not available from the second candidate block (RB b), motion vector candidate generation subroutine 700 proceeds to decision block 732.

  If, at decision block 732, a motion vector is available from the third candidate reference block (RBc), then at execution block 744, the motion vector candidate generation subroutine 700 converts the second motion vector candidate (MVC2) to an MV. Set to c. In an implementation block 740, the third motion vector candidate (MVC3) is set to (0,0).

  Next, the motion vector candidate generation subroutine 700 returns a motion vector candidate list having respective values of MVC 1 = MVa, MVC 2 = MVc, and MVC 3 = (0,0) in a return block 799.

  Referring again to decision block 732, if a motion vector is not available from the third candidate reference block (RB c), then at execution block 748, the motion vector candidate generation subroutine 700 proceeds to the second motion vector candidate (MVC 2). ) Is set to (0,0), and the third motion vector candidate (MVC 3) is set to (0,0) in the execution block 740.

  Next, in the return block 799, the motion vector candidate generation subroutine 700 generates a motion vector candidate list having respective values of MVC 1 = MVa, MVC 2 = (0,0), and MVC 3 = (0,0). return it.

  Referring again to decision block 708, if a motion vector is not available from the first candidate reference block (RBa), motion vector candidate generation subroutine 700 proceeds to decision block 720.

  If, at decision block 720, a motion vector is available from the second candidate reference block (RB b), then at implementation block 752, the motion vector candidate generation subroutine 700 converts the first motion vector candidate (MVC 1) to an MV Set to b. Next, the motion vector candidate generation subroutine 700 proceeds to decision block 732.

  Returning to decision block 732, if a motion vector is available from the third candidate reference block (RBc), then at execution block 744, the motion vector candidate generation subroutine 700 proceeds to the second motion vector candidate (MVC 2). Is set to MV c. In an implementation block 740, the third motion vector candidate (MVC3) is set to (0,0).

  Next, the motion vector candidate generation subroutine 700 returns, in a return block 799, a motion vector candidate list having respective values of MVC1 = MVb, MVC2 = MVc, and MVC3 = (0,0).

  Referring again to decision block 732, if a motion vector is not available from the third candidate reference block (RB c), then at execution block 748, the motion vector candidate generation subroutine 700 proceeds to the second motion vector candidate (MVC 2). ) Is set to (0,0), and in the execution block 740, the third motion vector candidate (MVC3) is set to (0,0).

  Next, the motion vector candidate generation subroutine 700 generates a motion vector candidate list having respective values of MVC 1 = MVb, MVC 2 = (0,0) and MVC 3 = (0,0) in a return block 799. return.

  Referring again to decision block 720, if a motion vector is not available from the second candidate reference block (RB b), motion vector candidate generation subroutine 700 proceeds to decision block 756.

  If, at decision block 756, a motion vector is available from the third candidate reference block (RBc), at execution block 760, the motion vector candidate generation subroutine 700 converts the first motion vector candidate (MVC 1) to the MV Set to c. Next, the motion vector candidate generation subroutine 700 sets the second motion vector candidate (MVC 2) to (0, 0) in the execution block 748, and sets the third motion vector candidate (MVC 3) in the execution block 740. ) To (0,0).

  Next, in the return block 799, the motion vector candidate generation subroutine 700 generates a motion vector candidate list having respective values of MVC 1 = MVc, MVC 2 = (0,0), and MVC 3 = (0,0). return it.

  Referring again to decision block 756, if a motion vector is not available from the third candidate reference block (RBc), then at implementation block 764, the motion vector candidate generation subroutine 700 proceeds to the first motion vector candidate (MVC 1). ) To (0,0). Next, the motion vector candidate generation subroutine 700 sets the second motion vector candidate to (0,0) in the execution block 748, and sets the third motion vector candidate to (0,0) in the execution block 740. Set.

  Next, in the return block 799, the motion vector candidate generation subroutine 700 generates a motion vector candidate list having respective values of MVC 1 = MVb, MVC 2 = (0,0), and MVC 3 = (0,0). return it.

Motion vector restoration routine 800
FIG. 8 shows a motion vector restoration routine 800 suitable for use with at least one embodiment such as the decoder 500. As will be appreciated by those skilled in the art, not all events in the decoding process are shown in FIG. Rather, for clarity, only those steps that are reasonably relevant to describing the motion vector selection routine are shown.

  In an implementation block 803, the motion vector restoration routine 800 obtains data corresponding to an encoded block.

At implementation block 828, the motion vector recovery routine 800 identifies the coding mode used to code the coded block. As described above, the encoding modes that may have been used are the inter encoding mode , the direct encoding mode, or the skip encoding mode.

If, at decision block 830, the coded block was coded using the inter-coding mode , then at implementation block 833, the motion vector restoration routine 800 identifies a predicted block corresponding to the coded block. In a start loop block 836, each prediction block in the current coded block is processed in turn.

  In an implementation block 839, the motion vector restoration routine 800 determines the prediction index of the current prediction block from the prediction block header.

  In implementation block 842, the motion vector recovery routine 800 may include a motion vector prediction method used to predict the motion vector of the current prediction block, for example, by reading the SMV-PM flag in the picture header of the current frame. To identify.

  In an implementation block 848, the motion vector restoration routine 800 obtains a motion vector difference (ΔMV) of the current prediction block.

  In an implementation block 851, the motion vector restoration routine 800 obtains a predicted motion vector (MV pred) of the current prediction block using the motion vector prediction method specified in the implementation block 842.

  In an implementation block 854, the motion vector restoration routine 800 may calculate the motion vector (MV calc) for the current predicted block by adding the predicted motion vector (MV pred) to the motion vector difference (ΔMV), for example. (For a P-type picture, there is a single restored motion vector, and for a B-type picture, there are two restored motion vectors).

  Next, in an implementation block 857, the motion vector restoration routine 800 adds the residual of the current prediction block to the block indicated by the calculated motion vector (MV calc) to obtain a restoration value of the prediction block. .

  Referring again to decision block 830, if the current coded block was coded using either the skip coding mode or the direct coding mode, the motion vector reconstruction routine 800 In response, a motion vector candidate generation subroutine 700 (described above with reference to FIG. 7) that returns an ordered list of motion vector candidates is called.

  Next, in an implementation block 863, the motion vector restoration routine 800 reads a motion vector selection flag from the prediction block header of the implementation block 863.

  Next, in an implementation block 866, the motion vector restoration routine 800 uses the motion vector selection flag to determine the motion vector from the ordered list of ordered motion vector candidate lists used to encode the current prediction block. To identify.

  If, at decision block 869, the current coded block was coded in the direct coding mode, then at implementation block 872, the motion vector restoration routine 800 uses the selected motion vector to recover the predicted block coefficients. The residual of the prediction block is added to the coefficient of the specified block.

  If the current coded block is coded in the skip coding mode, in an implementation block 875, the motion vector restoration routine 800 uses the coefficients of the reference block indicated by the selected motion vector as prediction block coefficients.

  The motion vector restoration routine 800 ends at end block 899.

  Although specific embodiments have been illustrated and described herein, those skilled in the art will be able to substitute and / or equivalently substitute for the specific embodiments shown or described. You can recognize that there is. This application is intended to cover any adaptations or variations of the embodiments discussed herein.

Claims (23)

A method of encoding a video frame in a sequence of video frames to generate an encoded bitstream representing an unencoded video frame, the method comprising:
Obtaining a coded block from the uncoded video frame;
Selecting one encoding mode from a plurality of encoding modes, including an inter encoding mode , a skip encoding mode, and a direct encoding mode, for use in encoding the encoded block;
With
When the inter-coding mode is selected,
Dividing the obtained encoded block into a plurality of prediction blocks;
For each prediction block of the plurality of prediction blocks,
Selecting one motion vector prediction method from among the motion vector prediction method using an intermediate value and the motion vector prediction method using an average value , for use in encoding each of the prediction blocks;
To generate a video data portion of each of said prediction block in the encoded bit stream, a step of coding each said prediction block based on motion-out vector prediction method selected,
Setting a selected motion vector prediction method flag indicating the selected motion vector prediction method used for encoding the prediction block in a header portion of each of the encoded prediction blocks ;
With
When the skip encoding mode or the direct encoding mode is selected, further,
Setting the obtained encoded block as a current prediction block;
Obtaining an ordered list of motion vector candidates for the current prediction block;
Selecting a motion vector from the ordered list of candidate motion vectors for use in encoding the current prediction block;
To generate a portion of the video data portion of the current prediction block in the encoded bit stream, a step of coding the current prediction block using-out the selected motion vector,
The encoded in the header of the current prediction block, and setting the motion vector selection flag indicating the position of the ordered list of the motion vector candidate corresponding to the moving-out vector said selected
With
Setting a coding mode selection flag indicating a coding mode selected from the plurality of coding modes for coding the coding block in a header portion of the coded bit stream. Method. When the skip encoding mode or the direct encoding mode is selected, further,
A plurality of prediction blocks arranged in a plurality of rows and a plurality of columns, wherein the current prediction block is located in a first row of the plurality of rows; Splitting into a plurality of prediction blocks located in a first column of the column,
Obtaining the ordered list of motion vector candidates,
Selecting a first motion vector used when previously encoding a first reference prediction block of the plurality of prediction blocks;
Selecting a second motion vector used when previously encoding a second reference prediction block of the plurality of prediction blocks;
Selecting a third motion vector used when previously encoding a third reference prediction block of the plurality of prediction blocks;
The method of claim 1, comprising: The first reference prediction block is located at a second row adjacent to the first row in the plurality of rows of the plurality of prediction blocks, and a prediction block located at the first column. Selected,
The second reference prediction block is located at a second row of the plurality of prediction blocks and a prediction block located at a second column adjacent to the first column in the plurality of columns. Selected,
The third reference prediction block is located at the first row of the plurality of prediction blocks, and the third reference prediction block is located at a third column adjacent to the first column in the plurality of columns. 3. The method of claim 2, wherein said method is selected. When the skip encoding mode or the direct encoding mode is selected, further,
A plurality of prediction blocks arranged in a plurality of rows and a plurality of columns, wherein the current prediction block is located in a first row of the plurality of rows; Splitting into a plurality of prediction blocks located in a first column of the column,
Obtaining the ordered list of motion vector candidates,
Selecting a first motion vector used when previously encoding a first reference prediction block of the plurality of prediction blocks;
Selecting a second motion vector used when previously encoding a second reference prediction block of the plurality of prediction blocks;
Selecting a zero value motion vector;
The method of claim 1, comprising: The first reference prediction block is located at a second row adjacent to the first row in the plurality of rows of the plurality of prediction blocks, and a prediction block located at the first column. Selected,
The second reference prediction block is located at a second row of the plurality of prediction blocks and a prediction block located at a second column adjacent to the first column in the plurality of columns. 5. The method according to claim 4, wherein the method is selected. The first reference prediction block is located at a second row adjacent to the first row in the plurality of rows of the plurality of prediction blocks, and a prediction block located at the first column. Selected,
The second reference prediction block is located at the first row of the plurality of prediction blocks, and the second reference prediction block is located at a second column adjacent to the first column in the plurality of columns. 5. The method according to claim 4, wherein the method is selected. The first reference prediction block is located in a second row of the plurality of prediction blocks adjacent to the first row in the plurality of rows, and is located in the first column in the plurality of columns. Selecting from a prediction block located in an adjacent second column,
The second reference prediction block is located at a first row of the plurality of prediction blocks and a prediction block located at a third column adjacent to the first column in the plurality of columns. 5. The method according to claim 4, wherein the method is selected. When the skip encoding mode or the direct encoding mode is selected, further,
A plurality of prediction blocks arranged in a plurality of rows and a plurality of columns, wherein the current prediction block is located in a first row of the plurality of rows; Splitting into a plurality of prediction blocks located in a first column of the column,
The ordered list of motion vector candidates is:
A first motion vector used when previously encoding a first reference prediction block of the plurality of prediction blocks;
A first zero-valued motion vector;
A second zero-valued motion vector;
The method of claim 1, comprising:   9. The method of claim 8, wherein the first reference prediction block is located in a second row adjacent to the first row in the plurality of rows and is located in the first column.   The first reference prediction block is located in a second row adjacent to the first row in the plurality of rows, and is located in a second column adjacent to the first column in the plurality of columns. 9. The method of claim 8, wherein the method comprises:   9. The method of claim 8, wherein the first reference prediction block is located in the first row and in a second column of the plurality of columns adjacent to the first column. The ordered list of motion vector candidates is:
A first zero-valued motion vector;
A second zero-valued motion vector;
A third zero value motion vector;
The method of claim 1, comprising: The step of encoding the current prediction block using said-out selected motion vectors,
A step of using the selected motion-out vector, identifying a second prediction block in the frame that is previously encoded in the sequence of video frames,
Obtaining a residual between the current prediction block and the second prediction block;
Encoding the residual to generate a portion of the video data portion of the encoded bitstream that corresponds to the current prediction block;
The method of claim 1, comprising: The step of encoding the current prediction block using said-out selected motion vectors,
A step of using the selected motion-out vector, identifying a second prediction block in the frame that is previously encoded in the sequence of video frames,
Encoding the second prediction block to generate a portion of the video data portion of the encoded bitstream that corresponds to the current prediction block;
The method of claim 1, comprising: A method of encoding a video frame to generate an encoded bitstream representing an unencoded video frame, the encoded bitstream including at least a header portion and a video data portion.
Obtaining a coded block from the uncoded video frame;
Selecting one encoding mode from a plurality of encoding modes, including an inter encoding mode , a skip encoding mode, and a direct encoding mode, for use in encoding the encoded block;
With
When the inter-coding mode is selected,
Dividing the obtained encoded block into a plurality of prediction blocks;
Selecting one motion vector prediction method from among the motion vector prediction method using an intermediate value and the motion vector prediction method using an average value, for use in encoding each of the plurality of prediction blocks,
For each of said plurality of prediction blocks, a step of acquiring the respective motion vector by using the selected motion-out vector prediction method,
Encoding each of the plurality of prediction blocks using each of the motion vectors to generate a portion of a video data portion of the encoded bitstream;
A selected motion vector prediction method flag indicating the selected motion vector prediction method used for encoding the plurality of prediction blocks in a header portion corresponding to the plurality of prediction blocks in the encoded bit stream. Setting the
Setting a coding mode selection flag indicating a coding mode selected from the plurality of coding modes in order to code the coding block, in a header portion of the coded bit stream;
A method comprising: At least one motion vector prediction method among the plurality of motion vector prediction methods,
A first motion vector used when previously encoding a first reference prediction block in the plurality of prediction blocks;
A second motion vector used when previously encoding a second reference prediction block in the plurality of prediction blocks;
A third motion vector used when previously encoding a third reference prediction block in the plurality of prediction blocks;
The method of claim 15, comprising determining an intermediate value of: At least one motion vector prediction method among the plurality of motion vector prediction methods,
A first motion vector used when previously encoding a first reference prediction block in the plurality of prediction blocks;
A second motion vector used when previously encoding a second reference prediction block in the plurality of prediction blocks;
16. The method of claim 15, comprising determining an average value of. When the skip encoding mode or the direct encoding mode is selected, further,
Setting the obtained encoded block as a current prediction block;
Obtaining an ordered list of motion vector candidates for the current prediction block;
Selecting a motion vector from the ordered list of candidate motion vectors for use in encoding the current prediction block;
Encoding the current prediction block using the motion vector to generate a portion of the video data portion of the current prediction block in the encoded bitstream;
Setting a motion vector selection flag indicating a position in an ordered list of the motion vector candidates corresponding to the motion vector in a header portion of the encoded current prediction block;
The method of claim 15 comprising: A method of encoding a video frame in a sequence of video frames to generate an encoded bitstream representing an unencoded video frame, the method comprising:
Obtaining a coded block from the uncoded video frame;
Selecting one encoding mode from a plurality of encoding modes, including an inter encoding mode , a skip encoding mode, and a direct encoding mode, for use in encoding the encoded block;
Setting a coding mode selection flag indicating the selected coding mode in a header portion of the coded bit stream;
Selecting a prediction block from the encoded block;
With
When the inter-coding mode is selected,
Selecting one motion vector prediction method from a motion vector prediction method using an intermediate value and a motion vector prediction method using an average value ;
Encoding a prediction block based on the motion vector prediction method selected to generate a video data portion in the encoded bitstream;
Setting a selected motion vector prediction method flag indicating a selected motion vector prediction method in a header portion of the encoded bit stream;
With
When the skip encoding mode or the direct encoding mode is selected, further,
To generate a video data portion of the encoded bit stream, a step of encoding the prediction block using a motion vector selected from an ordered list of candidate motion vectors,
Setting a motion vector selection flag indicating a position in the ordered list of the motion vector candidate corresponding to the motion vector in a header portion of the encoded prediction block;
A method comprising: Encoding the prediction block using the motion vector,
Obtaining an ordered list of the motion vector candidates;
Selecting the motion vector from the ordered list of motion vector candidates;
Using the motion vector to identify a second predicted block in a previously encoded frame in the sequence of video frames;
Obtaining a residual between the prediction block and the second prediction block;
Encoding the residual to generate a video data portion of the encoded bitstream;
20. The method of claim 19, comprising: A method of encoding a video frame in a sequence of video frames to generate an encoded bitstream representing an unencoded video frame, the method comprising:
Obtaining a coded block from the uncoded video frame;
Selecting one encoding mode from a plurality of encoding modes, including an inter encoding mode, a skip encoding mode, and a direct encoding mode, for use in encoding the encoded block;
With
When the inter-coding mode is selected,
Dividing the obtained encoded block into a plurality of prediction blocks;
For each prediction block of the plurality of prediction blocks,
Selecting one motion vector prediction method from among the motion vector prediction method using an intermediate value and the motion vector prediction method using an average value, for use in encoding each of the prediction blocks;
Encoding each of the prediction blocks based on a selected motion vector prediction method to generate a video data portion of each of the prediction blocks in the encoded bitstream;
Setting a selected motion vector prediction method flag indicating the selected motion vector prediction method used for encoding the prediction block in a header portion of each of the encoded prediction blocks;
With
When the skip encoding mode or the direct encoding mode is selected, further,
Setting the obtained encoded block as a current prediction block;
Obtaining an ordered list of motion vector candidates for the current prediction block,
Selecting a first motion vector used in previously encoding a first reference prediction block of a plurality of prediction blocks for the current prediction block;
Selecting a second motion vector used when previously encoding a second reference prediction block of the plurality of prediction blocks;
Selecting a third motion vector used when previously encoding a third reference prediction block of the plurality of prediction blocks;
A process comprising:
Selecting a motion vector from the ordered list of candidate motion vectors for use in encoding the current prediction block;
Encoding the current prediction block using a selected motion vector to generate a portion of the video data portion of the current prediction block in the encoded bitstream;
Setting a motion vector selection flag indicating a position in an ordered list of the motion vector candidates corresponding to the selected motion vector in a header portion of the encoded current prediction block;
With
Setting a coding mode selection flag indicating a coding mode selected from the plurality of coding modes for coding the coding block in a header portion of the coded bit stream. Method. A method of encoding a video frame in a sequence of video frames to generate an encoded bitstream representing an unencoded video frame, the method comprising:
Obtaining a coded block from the uncoded video frame;
Selecting one encoding mode from a plurality of encoding modes, including an inter encoding mode, a skip encoding mode, and a direct encoding mode, for use in encoding the encoded block;
With
When the inter-coding mode is selected,
Dividing the obtained encoded block into a plurality of prediction blocks;
For each prediction block of the plurality of prediction blocks,
Selecting one motion vector prediction method from among the motion vector prediction method using an intermediate value and the motion vector prediction method using an average value, for use in encoding each of the prediction blocks;
Encoding each of the prediction blocks based on a selected motion vector prediction method to generate a video data portion of each of the prediction blocks in the encoded bitstream;
Setting a selected motion vector prediction method flag indicating the selected motion vector prediction method used for encoding the prediction block in a header portion of each of the encoded prediction blocks;
With
When the skip encoding mode or the direct encoding mode is selected, further,
Setting the obtained encoded block as a current prediction block;
Obtaining an ordered list of motion vector candidates for the current prediction block,
Selecting a first motion vector used in previously encoding a first reference prediction block of a plurality of prediction blocks for the current prediction block;
Selecting a second motion vector used when previously encoding a second reference prediction block of the plurality of prediction blocks;
Selecting a zero value motion vector;
A process comprising:
Selecting a motion vector from the ordered list of candidate motion vectors for use in encoding the current prediction block;
Encoding the current prediction block using a selected motion vector to generate a portion of the video data portion of the current prediction block in the encoded bitstream;
Setting a motion vector selection flag indicating a position in an ordered list of the motion vector candidates corresponding to the selected motion vector in a header portion of the encoded current prediction block;
With
Setting a coding mode selection flag indicating a coding mode selected from the plurality of coding modes for coding the coding block in a header portion of the coded bit stream. Method. A method of encoding a video frame in a sequence of video frames to generate an encoded bitstream representing an unencoded video frame, the method comprising:
Obtaining a coded block from the uncoded video frame;
Selecting one encoding mode from a plurality of encoding modes, including an inter encoding mode, a skip encoding mode, and a direct encoding mode, for use in encoding the encoded block;
With
When the inter-coding mode is selected,
Dividing the obtained encoded block into a plurality of prediction blocks;
For each prediction block of the plurality of prediction blocks,
Selecting one motion vector prediction method from among the motion vector prediction method using an intermediate value and the motion vector prediction method using an average value, for use in encoding each of the prediction blocks;
Encoding each of the prediction blocks based on a selected motion vector prediction method to generate a video data portion of each of the prediction blocks in the encoded bitstream;
Setting a selected motion vector prediction method flag indicating the selected motion vector prediction method used for encoding the prediction block in a header portion of each of the encoded prediction blocks;
With
When the skip encoding mode or the direct encoding mode is selected, further,
Setting the obtained encoded block as a current prediction block;
An ordered list of motion vector candidates for the current prediction block,
A first motion vector used when previously encoding a first reference prediction block of the plurality of prediction blocks for the current prediction block;
A first zero-valued motion vector;
A second zero-valued motion vector;
Obtaining an ordered list containing
Selecting a motion vector from the ordered list of candidate motion vectors for use in encoding the current prediction block;
Encoding the current prediction block using a selected motion vector to generate a portion of the video data portion of the current prediction block in the encoded bitstream;
Setting a motion vector selection flag indicating a position in an ordered list of the motion vector candidates corresponding to the selected motion vector in a header portion of the encoded current prediction block;
With
Setting a coding mode selection flag indicating a coding mode selected from the plurality of coding modes for coding the coding block in a header portion of the coded bit stream. Method.

Images